Academia.eduAcademia.edu

Figure 4 – uploaded by Igor Zurbenko

See full PDFdownloadDownload figure
Fig. 3. Time series data from Chautauqua during 1984-1998: (a) raw precipitation-weighted SO; ~ ; (b) precipitation-weighted SO; ~ smoothed with the KZ;-y,, 3 filter; (c) precipitation-weighted SO; ~ smoothed with a KZA4.,,, 3 filter, depicting a large break around 1990. All data were first log-transformed to reduce nonlinear effects and stabilize the variance.

Figure 3 Time series data from Chautauqua during 1984-1998: (a) raw precipitation-weighted SO; ~ ; (b) precipitation-weighted SO; ~ smoothed with the KZ;-y,, 3 filter; (c) precipitation-weighted SO; ~ smoothed with a KZA4.,,, 3 filter, depicting a large break around 1990. All data were first log-transformed to reduce nonlinear effects and stabilize the variance.

Related Figures (9)

Summary of the Monte Carlo simulations to evaluate the performance of the KZA filter. The different cases represent deviations fro1 the base case of zero trend, seasonal variation of +30%, a short-term component consisting of a random sample from N(0, 1), and discontinuity of 50% at the beginning of year 11 in the 20-yr long time series
Summary of the Monte Carlo simulations to evaluate the performance of the KZA filter. The different cases represent deviations fro1 the base case of zero trend, seasonal variation of +30%, a short-term component consisting of a random sample from N(0, 1), and discontinuity of 50% at the beginning of year 11 in the 20-yr long time series
Fig. 1. Time series of raw (unfiltered) precipitation-weighted SO;~ and NO, data, and the corresponding filtered time series depicting the long-term components at the two collocated sites at Whiteface Mountain, NY, operated by the NADP and DEC.
Fig. 1. Time series of raw (unfiltered) precipitation-weighted SO;~ and NO, data, and the corresponding filtered time series depicting the long-term components at the two collocated sites at Whiteface Mountain, NY, operated by the NADP and DEC.
Fig. 2. Base case concentrations and percent improvements (normalized differences) from the base case resulting from the CAAA- mandated controls as simulated by RADM—SO) concentrations, and wet deposition of SO;~ and NO, . The base case simulations used 1990 emissions. climatological or longer-term variations (i.e., KZ}-yr,3, associated with time scales of > 1 yr) forcings to the total variance of the data. In the SO, and aerosol SO; ~ data, about 80-90% of the total variance is attributable to the components having time scales of <1 yr (synoptic, intra- seasonal, and seasonal variations), while long-term variability (trend component) accounts for only about 10-20% of the total variance. For precipitation- weighted SO; ~ , only about 2-15% of the total variance can be attributed to the long-term component, while for NO, less than 7% of the variability is attributable to the long-term components. Since much of the observed variability in these pollutant time series is attributable to natural fluctuations, the task of discerning changes in pollutant levels stemming from changes in emissions due to control policies is complicated. reduce ambient SQ» concentrations in the New York area of the modeling domain by about 20-30%, wet deposition SO; ~ by about 10-20%, and wet deposition NO, by about 10-20% in the Adirondacks and parts of western New York. As expected, the largest reduc- tions in SO, concentrations are predicted over portions of IL, IN, KY, OH, PA, and WV, where major point sources are located (Fig. 2). The reduction in the wet deposition of NO, of 10-20% was more uniform over much of the northeastern US. In the absence of natural variability, the RADM system predicted SO; reduc- tions ranging from 4% to 31%, and NO, reductions from 3% to 16% at the sites examined here, attributable to the implementation of Title IV of the CAAA (presented later in Table 3).
Fig. 2. Base case concentrations and percent improvements (normalized differences) from the base case resulting from the CAAA- mandated controls as simulated by RADM—SO) concentrations, and wet deposition of SO;~ and NO, . The base case simulations used 1990 emissions. climatological or longer-term variations (i.e., KZ}-yr,3, associated with time scales of > 1 yr) forcings to the total variance of the data. In the SO, and aerosol SO; ~ data, about 80-90% of the total variance is attributable to the components having time scales of <1 yr (synoptic, intra- seasonal, and seasonal variations), while long-term variability (trend component) accounts for only about 10-20% of the total variance. For precipitation- weighted SO; ~ , only about 2-15% of the total variance can be attributed to the long-term component, while for NO, less than 7% of the variability is attributable to the long-term components. Since much of the observed variability in these pollutant time series is attributable to natural fluctuations, the task of discerning changes in pollutant levels stemming from changes in emissions due to control policies is complicated. reduce ambient SQ» concentrations in the New York area of the modeling domain by about 20-30%, wet deposition SO; ~ by about 10-20%, and wet deposition NO, by about 10-20% in the Adirondacks and parts of western New York. As expected, the largest reduc- tions in SO, concentrations are predicted over portions of IL, IN, KY, OH, PA, and WV, where major point sources are located (Fig. 2). The reduction in the wet deposition of NO, of 10-20% was more uniform over much of the northeastern US. In the absence of natural variability, the RADM system predicted SO; reduc- tions ranging from 4% to 31%, and NO, reductions from 3% to 16% at the sites examined here, attributable to the implementation of Title IV of the CAAA (presented later in Table 3).
“Note: the uncertainty in break magnitude is about 12%. Negative values indicate a downward break and positive values indicate an upward break. The improvements predicted by RADM are included, as are the differences between the 1983-94 and 1995-97 averages. The time of occurrence (location) and magnitudes of breaks in SO, and aerosol SO; ~ time series® Table 2
“Note: the uncertainty in break magnitude is about 12%. Negative values indicate a downward break and positive values indicate an upward break. The improvements predicted by RADM are included, as are the differences between the 1983-94 and 1995-97 averages. The time of occurrence (location) and magnitudes of breaks in SO, and aerosol SO; ~ time series® Table 2
Fig. 4. Same as Fig. 3 except for data from Whiteface Mountain, NY.
Fig. 4. Same as Fig. 3 except for data from Whiteface Mountain, NY.
K.L. Civerolo et al. | Atmospheric Environment 35 (2001) 4135-4148 Fig. 5. Same as Fig. 3 except for data from Oxford, OH.
K.L. Civerolo et al. | Atmospheric Environment 35 (2001) 4135-4148 Fig. 5. Same as Fig. 3 except for data from Oxford, OH.
The time location and magnitudes of downward breaks in precipitation-weighted SO; and NO, , at the four NADP/DEC station pairs nearly collocated in space. Negative values indicate a downward break. The improvements predicted by RADM are also included. If a linear trend is more appropriate, the percent change, based on the slope of the linear trend, is shown
The time location and magnitudes of downward breaks in precipitation-weighted SO; and NO, , at the four NADP/DEC station pairs nearly collocated in space. Negative values indicate a downward break. The improvements predicted by RADM are also included. If a linear trend is more appropriate, the percent change, based on the slope of the linear trend, is shown
Fig. 6. The long-term components in precipitation-weighted NO; concentration data, exhibiting complex temporal structure: (a Chaswell, MI, (b) Wellston, MI, (c) SW Purdue Agricultural Center, IN, (d) Lilley Cornett Woods, KY, (e) Mackvile, KY, and (f Charlottesville, VA.
Fig. 6. The long-term components in precipitation-weighted NO; concentration data, exhibiting complex temporal structure: (a Chaswell, MI, (b) Wellston, MI, (c) SW Purdue Agricultural Center, IN, (d) Lilley Cornett Woods, KY, (e) Mackvile, KY, and (f Charlottesville, VA.
Related topics:
Environmental EngineeringAir QualityAtmospheric sciencesTrend EstimationAcid DepositionTrend detection
580 California St., Suite 400
San Francisco, CA, 94104
© 2025 Academia. All rights reserved